17936491. DATA SELECTION FOR AUTOMATED RETRAINING IN CASE OF DRIFTS IN ACTIVE LEARNING simplified abstract (International Business Machines Corporation)
Contents
DATA SELECTION FOR AUTOMATED RETRAINING IN CASE OF DRIFTS IN ACTIVE LEARNING
Organization Name
International Business Machines Corporation
Inventor(s)
Venkata Sitaramagiridharganesh Ganapavarapu of Elmsford NY (US)
Seshu Tirupathi of Dublin (IE)
DATA SELECTION FOR AUTOMATED RETRAINING IN CASE OF DRIFTS IN ACTIVE LEARNING - A simplified explanation of the abstract
This abstract first appeared for US patent application 17936491 titled 'DATA SELECTION FOR AUTOMATED RETRAINING IN CASE OF DRIFTS IN ACTIVE LEARNING
Simplified Explanation
The abstract describes a computer-implemented method for retraining a machine learning model in case of a drift. The method involves detecting a drift, identifying features and responses of the model, determining a time window of the drift, extracting data within the time window, checking if the data is sufficient for retraining, interpolating features and responses if necessary, and retraining the model using the interpolated data.
- Explanation of the patent:
- Detect drift in machine learning - Identify features and responses of the model - Determine time window of the drift - Extract data within the time window - Check if data is sufficient for retraining - Interpolate features and responses if needed - Retrain the model using the interpolated data
- Potential applications of this technology:
- Continuous improvement of machine learning models - Enhanced accuracy and performance of predictive models - Real-time adaptation to changing data patterns
- Problems solved by this technology:
- Addressing drift in machine learning models - Ensuring model accuracy over time - Automating retraining processes
- Benefits of this technology:
- Improved model performance - Reduced manual intervention for model maintenance - Increased efficiency in adapting to data changes
- Potential commercial applications of this technology:
- Predictive maintenance in manufacturing - Fraud detection in financial services - Personalized recommendations in e-commerce
- Possible prior art:
- Existing methods for retraining machine learning models - Techniques for detecting drift in data streams
- Unanswered Questions:
- 1. How does the method determine the sufficiency of extracted data for retraining the model?
- Unanswered Questions:
The abstract does not provide details on the specific criteria or algorithms used to determine if the extracted data is sufficient for retraining the model.
- 2. What interpolation techniques are employed for features and responses in case of insufficient data?
The abstract mentions interpolation of features and responses for a future time horizon but does not specify the interpolation methods or algorithms utilized in the retraining process.
Original Abstract Submitted
A computer-implemented method, a computer program product, and a computer system for retraining a model in case of a drift in machine learning. A computer detects a drift in machine learning. A computer identifies in a database features and a response of a machine learning model. A computer determines a time window of the drift. A computer extracts, from the database, data of the features and the response in the time window. A computer determines whether extracted data is sufficient for retraining the machine learning model. A computer, in response to determining that the extracted data is not sufficient for retraining the machine learning model, interpolates one or more of the features for a predetermined future time horizon. A computer interpolates a response corresponding to one or more interpolated features. A computer retrains the machine learning model, using the one or more interpolated features and an interpolated response corresponding thereto.
